Sturdevant et al.: Feeding habits, prey fields, and potential competition of Theragra chalcogrcimma and Clupea pallasi 
495 
son, 1994; Arrhenius, 1996; Haegele, 1997; Willette et al., 
1997). In addition to pelagic prey, epibenthic prey such as 
harpacticoid copepods and gammarid amphipods were im- 
portant in estuarine habitats of the Fraser River for Pacific 
herring and coastal Maine for Atlantic herring (Sherman 
and Perkins, 1971; Blaxter and Hunter, 1982; Levings, 
1983, in Lassuy, 1989). Spatial differences were found for 
diets of Pacific herring from four widespread bays in PWS 
(Foy and Norcross, 1999a). We found seasonal differences 
in prey similar to reports by others: euphausiids replaced 
calanoids in the diets of older juvenile herring compared 
with younger juvenile herring (Wailes,1936; Lassuy, 1989; 
Haegele, 1997), and larvaceans (Foy and Norcross, 1999a, 
1999b), mysids, and other malacostracans (Foy and Nor- 
cross, 1999a; Foy and Paul, 1999) were also more common 
in autumn. In Auke Bay, Alaska, juvenile herring diets 
varied with abundance of zooplankton prey taxa in spring 
and early summer and included large calanoids and eu- 
phausiids when they were present (Coyle and Paul, 1992). 
The marine distribution of the alewife, Alosa pseudoharen- 
gus, another herring, was correlated with the seasonal dis- 
tribution, availability, and abundance of its euphausiid 
prey (Stone and Jessop, 1994). 
Prey selection, feeding time, and depth 
Segregation by depth is one way to reduce interspecific 
competition among fish with overlapping distributions 
(e.g. Jessop, 1990; Arrhenius, 1996). Both pollock and her- 
ring perform diel vertical migrations (Smith, 1981; Blaxter 
and Hunter, 1982), the pattern of which can vary season- 
ally and ontogenetically (e.g., Kamba, 1977; Lassuy, 1989; 
Olla et al., 1996; Stokesbury et al., 2000). We observed 
some interspecific differences in prey selection that may 
relate to diel vertical distributions of predator and prey, 
prey preferences, ontogenetic changes in prey size, or dif- 
ferent feeding rhythms for both species. Neither species 
selected small calanoids in summer, and pollock selected 
taxa that herring did not (large calanoids, pteropods, and 
larvaceans). The summer herring were located nearshore 
and at the surface where densities of their main prey were 
twice as high as deep in the water column and where light 
for feeding was most intense. In contrast, summer pol- 
lock were located in relatively deep water with less light 
and lower prey densities. Even though small calanoids 
were the predominant prey of pollock and the predomi- 
nant zooplankter, they were avoided in relation to their 
availability. In autumn, both species avoided small cala- 
noids even more strongly than in summer. In autumn, 
sympatric pollock in shallow water selected large cala- 
noids, mainly Metrida pacifica, that were less available 
than in deeper water, but only the deeper allopatric her- 
ring selected them. Young Pacific herring in another study 
selected Calanus pacificus over Metridia lucens , perhaps 
in relation to fine-scale differences in prey depth distri- 
butions (Fortier and Leggett, 1983 in Munk et al., 1989; 
Bollens et al., 1993). Both sympatric herring and allo- 
patric herring strongly selected the larvaceans that were 
more evenly distributed in the water column. We primar- 
ily sampled fish schools that were located acoustically and 
assumed that the fish would be located where food was 
available (e.g. Arrhenius and Hansson, 1999). Compari- 
sons of prey selection between seasons are valid because 
the calculations are based on prey percentages which did 
not differ between the 243- and 303-pm mesh sizes used. 8 
The less-digested condition of herring stomach contents 
by day compared with night suggests that the summer 
fish and autumn allopatric fish were actively feeding and 
that the autumn sympatric fish were not. The condition 
of pollock prey and the fact that the fish were not located 
where their selected prey were in either season suggest 
that pollock were not actively feeding at the times we sam- 
pled in either season. The change in digested state of their 
prey with time of day also suggests that YOY pollock may 
seasonally or ontogenetically switch from feeding princi- 
pally during the day to feeding at night 12 (Merati and Bro- 
deur, 1996; Brodeur et al., 2000). 
Past diel studies have reported different patterns of 
feeding for pollock and various herring species. For exam- 
ple, peak time of feeding for pollock was at midnight or just 
before dawn in some studies (Brodeur and Wilson, 1996a; 
Merati and Brodeur, 1996; Willette et al., 1997), but an- 
other study showed a change in feeding chronology from 
sunset in small fish to night in larger fish (Brodeur et 
al., 2000). Peak time of feeding for Atlantic, Baltic, and 
Pacific herring occurred in the afternoon or evening, with 
the lowest feeding rates in early morning (e.g. de Silva, 
1973; Blaxter and Hunter, 1982; Mehner and Heerkloss, 
1994; Willette et al., 1997; Arrhenius, 1998). Changes in 
prey composition with time of day and ontogeny have also 
been noted in some studies (e.g. Nakatani, 1988; Grover, 
1991; Munk, 1992; Stone and Jessop, 1994) and not others 
(Brodeur et al., 2000). If co-occurring fish feed at different 
times, their diets could be highly similar without direct 
competition because predation on the same prey resourc- 
es would be temporally or spatially separated. This could 
be the case with small calanoid prey, which both species 
fed on in summer, but only herring were co-located at the 
depth of prey concentration. Different feeding periodicities 
could result in indirect competition if prey resources are 
limited, however. 
Sampling time could also affect the appearance in the 
diet of vertically migrating macrozooplankton, such as 
juvenile-adult euphausiids and hyperiid amphipods, be- 
cause the vertical locations of peak abundance of preda- 
tor and prey may not overlap continually. For example, 
euphausiids could be consumed at night near the surface 
or during the day near the bottom 12 (Pearcy et al., 1979; 
Nakatani, 1988). Juvenile euphausiids were rare summer 
prey in our study and were present only in those trawls 
taken below the 60-m mean depth of pollock catches. 2 In 
autumn, the euphausiid and fish distributions were more 
likely to overlap during the night sampling time (e.g. Bai- 
ley, 1989; Bollens et al., 1992a), and euphausiids were in- 
deed a principal prey in terms of biomass, particularly for 
pollock. Large calanoids and euphausiids could have been 
consumed at different feedings, particularly if their verti- 
cal distributions overlapped with those of the fish at dif- 
ferent times. Predation by pollock and herring on euphau- 
siids in areas where these macrozooplankters were not 
